Asixty-eight-year-old man who had been diagnosed with severe osteoarthritis
of the left knee presented for a total knee replacement. His medical history
was unremarkable, and he took no medications other than nonsteroidal
anti-inflammatory drugs. There was no family history of clotting abnormalities
or stroke. The preoperative laboratory evaluation did not reveal any
abnormality; the prothrombin time, activated partial thromboplastin time,
international normalized ratio, and platelet count were all within normal
limits. An epidural catheter was inserted prior to the operation to provide
anesthesia during surgery and analgesia postoperatively. According to the
medical record, the catheter was inserted without difficulty, the operation
was completed uneventfully, and the patient was then transferred to a ward.
Low-molecular-weight heparin (tinzaparin sodium, 4500 anti-factor Xa IU) was
injected subcutaneously into the abdominal wall six hours after the operation
and then once daily. Blood loss from the drain in the knee joint was 420 mL,
which was considered unremarkable, and the drain was removed on the second
postoperative day. Analgesia was provided through the epidural catheter with
Naropein (ropivacaine hydrochloride) 0.1% (AstraZeneca, Monts, France) and
fentanyl 2 mg/mL at an administration rate of 7 mL/h. Mobilization of the
patient began on the first postoperative day. The epidural catheter was
removed on the third postoperative day, twelve hours after administration of
the anticoagulant injection, and the patient was allowed to have oral
analgesics as needed. The next day (four days postoperatively), the patient
reported muscle weakness and low-back pain. Neurologic evaluation revealed
muscle weakness in all muscle groups of the lower extremities. According to
the grading scale of the British Medical Research
Council2 (in which a
grade of 5 indicates normal strength), muscle strength was graded as 4 for the
iliopsoas, 3 for the quadriceps, 4 for the hamstrings, 4 for the anterior
tibial muscles, and 4 for the posterior tibial muscles. An epidural hematoma
was suspected, and the low-molecular-weight heparin was discontinued. Until
that time, the patient had received one dose of tinzaparin each day for four
days. The last dose (fourth dose) was given on the third postoperative day,
which was the day of the catheter removal.
An urgent magnetic resonance imaging scan was ordered, but the scanner was
not available. In the afternoon, a clinical evaluation revealed increased
muscle weakness in all muscle groups of the lower extremities. The muscle
strength was rated as 3 in all muscle groups. The next morning, the patient
demonstrated further weakening in the lower extremities, and muscle strength
in all muscle groups was rated as 2 (iliopsoas, quadriceps, and hamstrings
muscles) or 3 (anterior tibial and posterior tibial muscles). The patient had
also lost control of bowel and bladder sphincters. A magnetic resonance
imaging scan was acquired that morning (twenty-six hours after the first
symptoms) and revealed an epidural hematoma from the level of T12 to L3
(Fig. 1). The prothrombin time,
activated partial thromboplastin time, international normalized ratio, and
platelet counts were still within normal limits.
The patient was taken to the operating room approximately thirty-two hours
after the symptoms began, and an emergent laminectomy at the level of the
hematoma was performed. Two laminae (L1 and L2) were resected without
disturbing the facets, and the vertebral canal was explored. An extensive clot
had formed and was compressing the dura and its contents. The hematoma was
meticulously evacuated, but no actively bleeding vessels were identified. The
wound was closed. In the recovery room, a bolus dose of 1000 µg of
methylprednisolone (Solu-Medrol; Pfizer Hellas, Athens, Greece) was
administered to the patient over the course of one hour. No additional doses
of methylprednisolone were given.
Subsequently, the patient remained in bed for two days and was evaluated
regularly, but his neurological status did not improve during that time. The
third day after the laminectomy, the patient started to show neurological
improvement, with increased muscle strength in the anterior and posterior
tibial muscles. Improvement continued, and, on the eighth day after the
laminectomy, the patient could move his legs freely in bed. On the twelfth day
after the laminectomy, he could stand alone and take a few steps with a walker
and, according to the grading scale of the British Medical Research Council,
muscle strength was rated as 4 for both iliopsoas muscles, 5 for the right
quadriceps and 4 for the left quadriceps, and 4 for the anterior and posterior
tibial muscles in both legs. The patient was discharged from the hospital to a
rehabilitation center on the twelfth day after the laminectomy and was
reevaluated three weeks later (thirty-three days after the laminectomy). At
that time, he had no residual neurological deficit, and the range of motion in
the involved knee ranged from full extension to 95° of flexion. No further
special coagulation studies were performed.
The patient underwent the usual follow-up for a patient who had undergone
joint replacement. At one year after knee replacement, the patient had normal
muscle strength, and the range of motion in the knee ranged from full
extension to 100° of flexion. At the time of the two-year postoperative
followup, the results were similar to those obtained at one year.
The rate of bleeding complications associated with epidural anesthesia is
estimated to range from 1 in every
150,0003 to 1 in
every
1,000,0004,5
patients. Although rare, this complication may be catastrophic as it can
result in a severe neurologic deficit, such as paraplegia. Vandermeulen et al.
identified two major risk
factors6. The first
factor is difficult or traumatic insertion or removal of the catheter
(including multiple punctures), and the second is abnormalities of the
coagulation mechanism. These abnormalities can be due to conditions that
existed preoperatively (such as thrombocytopenia, chronic renal failure, or
chronic liver disease) or can be caused by anticoagulant medications that are
administered to the patient to prevent deep venous thrombosis. As reported in
the literature, administration of heparin or low-molecular-weight heparin is a
recognized factor for the development of an epidural hematoma, especially when
the epidural catheter was introduced with notable
difficulty7. Intake
of salicylates may be associated with an increased risk of epidural bleeding
and subsequent hematoma, but heparin and low-molecular-weight heparin are even
more strongly associated with that
risk8. In our review
of the literature, we did not find a case of epidural hematoma in a patient
receiving Coumadin (warfarin) who had an international normalized ratio of
=2.0. The source of the hematoma is believed to be an injured epidural
artery or a vein from the epidural venous
plexus4. It is well
known that anticoagulant treatment is associated with hematoma
formation7,8.
Although there are certain instructions regarding the administration of
anticoagulants and the removal of the spinal
catheters9, the
possibility of hematoma formation cannot be excluded. Early diagnosis and
treatment of a hematoma is crucial to avoid permanent neurologic deficits. A
high index of suspicion should be maintained and careful and repeated clinical
examinations are necessary to assess symptoms, especially if there is gradual
worsening. Magnetic resonance imaging is the preferred modality to confirm
compression of the spinal cord and nerve
roots10. Immediate
laminectomy and surgical evacuation of the hematoma is indicated to decompress
the spinal cord. The surgeon must be aware that, even when early action has
been taken, the outcome may not be
optimal11.
Although the literature contains many case reports of epidural hematoma
after removal of an epidural
catheter12-15,
we are aware of only one report describing an epidural hematoma that occurred
after spinal anesthesia in total knee
arthroplasty16. In
that report, the epidural hematoma occurred immediately after surgery and
sensorimotor disturbances were present before removal of the catheter.
Nonoperative treatment was chosen, and sensorimotor recovery was complete six
months after surgery.
In our patient, the hematoma occurred after removal of the epidural
catheter three days after the operation. In our opinion, there was not an
obvious reason to explain the formation of the epidural hematoma. Insertion of
the catheter was not difficult, and it was removed twelve hours after the last
administration of low-molecular-weight heparin in accordance with
recommendations for safe removal of
catheters9,17.
We think the likely cause for the hematoma formation was an arterial or venous
injury that occurred in the vertebral canal during removal of the catheter.
Surgical decompression of the spinal canal with laminectomy and evacuation of
the hematoma was the chosen course of action, and the patient fully recovered.
Fortunately, the delay in diagnosis for this patient did not have serious
consequences.
Although rare, hematoma formation should be suspected when a patient
presents with symptoms suggesting spinal cord compression after epidural or
spinal anesthesia and the administration of low-molecular-weight heparin.
Immediate intervention to evacuate the hematoma and decompress the vertebral
canal must be undertaken to optimize neurological recovery.